REFRACTORIES amorphous states, the point of optical deorientation, the break in electrical conductivity, the discontinuity of the heating curve, the change in specific volume, etc. For this reason plots showing the relation between specific gravity and temperature frequently establish critical points of considerable interest. At the same time it might be understood that even the most refined methods of the laboratory will fail in dealing with systems of great viscosity in which the high internal friction may arrest or greatly delay any changes by means of which the critical point is recognized. Resistance to Compression at Furnace Temperatures.—In use refractories are frequently subjected to compression and sometimes tension stresses. This is particularly the case in the arches and crowns of furnaces and kilns, gas-retort benches, hot-blast stoves, etc. Any decrease in. rigidity suffered by the refractory body due to incipient softening will at once become evident by more or less marked deformation under stresses exceedingly small compared with the eompres-sive strengths of the material in the cold state. Thus, a firebrick showing a crushing strength in the cold of 3,000 Ib. per square inch may be deformed under a pressure of 50 Ib. per square inch at furnace temperatures. The softening of such substances through a given viscosity range is not only a function of temperature but of time as well. It is possible therefore to vitrify a body by exposure to higher temperatures in a shorter time and at lower points of the temperature scale during longer periods. Great differences exist as to the viscosity and deformation of different minerals and rocks both natural and artificial. As a general proposition it may be stated that the higher the material is in fluxing impurities, i.e. basic oxides in the case of siliceous and acid oxides in basic mixtures, the greater the deformation under conditions of pressure. It is a fact also that silicates high in alumina show greater contraction tinder constant pressure than siliceous ones. The resistance of refractories to pressure at furnace temperatures therefore offers a measure of the rigidity of refractories and, indeed, of their resistance to heat, since it is evident that a material showing but little deformation at a given temperature, say 1,350°C., and under a given load is to be preferred to one which softens and gives evidence of excessive contraction. With reference to resistance to load conditions it may be said that the refractoriness of a material is a function of the pressure employed. For instance, a firebrick, showing in the unloaded condition a softening point corresponding to cone 32 to 33, may under pressure show a steady deformation or collapse at a temperature represented by cone 19 to 20.l The initial mechanical strength and the degree of firing are, however, factors in determining the resistance to load conditions, since dense brick, formed under a heavy pressure,-are more resistant than those of a more open and porous structure. The amount of cementing material present and the degree in which it coheres with the granular constituents is involved as well. Almost invariably hard-fired bricks in which this condition is brought to a high degree of perfection will stand up much better than those burned to a lower temperature. It will be seen therefore that resistance to pressure at furnace temperature is by no means a simple property and a test involving such a condition should not be prescribed indiscriminately. It should be reserved for refractories which are actually to be used in places where pressure .is really an important factor. The test as applied so -far in this country errs on the safe side but is liable to be unfair to certain materials. This defect is diminished by reducing the load applied and for this reason it is becoming customary to employ a pressure of 25 Ib. per square inch rather than 50 Ib. as has been the former practice. Failure may occur in one of two ways, either through the softening of the mass as evidenced by excessive contraction, distortion or bending or through shear, in which i J. W. MELLOK and W. EMBPY, Trans. English Ceramic Soc., Vol. 17, part 2, pp. 360 et $eq.